I want to remove _custom_exporter from all the metrics being present on the Grafana dashboard.
How do i put this in legend format?
i should display only OS names.
like ubuntu20_04 instead of ubuntu20_04_custom_exporter.
Can you please help me?
Hi @murarisaranvikas, welcome to the community!
This post in the community has a lot of great information about how to use the value mapping feature with regex to edit values that show up in your graph.
I think we could definitely use more examples, so I mocked up your query with CSV data and
a bar chart to show an example of how you could remove the end of the value name in your bar chart.
To transform this:
Into this, removing the _exporter string at the end:
You can use a value mapping in the bar gauge options in the sidebar of the panel editor.
When you click Add value mappings it opens with a default Value mapping not Regex.
Click the Add a value mapping button, and choose Regex from the dropdown to add the regex value mappng and substitution. You can delete the default value mapping.
The regex (.*)_.* will capture everything up to the last underscore and $1 will substitute that as the new value.
The site https://regex101.com/ is a really helpful tool to test out regex. You’ll want to change the flavor to golang to get the right interpretations for Grafana.
5 Likes
Hi there. Thanks for the great explanation! Unfortunately I have a similar problem but the regex doesn’t work. I need to get the last part of the string as legend entry. Although it is working great on https://regex101.com/, it is still not working in Grafana:
What I am doing wrong?
Thanks ans regards
JLER
This works for me if I use the Transform option in the panel.
I am trying to do something similar, but remove a string before a character. Any thoughts on the regex?
Hi,
what is your data and what do want to achieve?
Regards,
Fadjar
We found a solution.
I wanted to remove the text after the last space:
some text the-rest-of-it essentially needed to keep the-rest-of-it and remove some text
we managed it using this transform:
1 Like